Models integrating genomic, environmental, and behavioral variables achieved an AUC of 0.87 for breast cancer and 0.83 for ovarian cancer, outperforming PRS-only models.
Cohort (n=5,128)
Yes
Does integrating environmental and lifestyle data with genomic risk scores improve predictive accuracy for breast and ovarian cancer in women aged 25-70?
Integrating environmental and lifestyle data with genomic risk scores significantly improves the predictive accuracy for breast and ovarian cancer compared to genomic data alone.
Absolute Event Rate: 0.87% vs 0.74%
p-value: p=<0.001
Objectives/Goals: To evaluate how integrating genomic risk scores with environmental pollutants, endocrine disruptor exposure, and lifestyle factors enhances predictive accuracy for breast and ovarian cancer, improving early-risk stratification and advancing equitable precision prevention among diverse female populations. Methods/Study Population: A secondary data analysis was conducted using publicly available datasets from the UK Biobank and NIH All of Us cohorts (n = 5,128 women aged 25–70). Polygenic risk scores (PRS) for breast and ovarian cancer were combined with environmental exposures (PM 2.5, endocrine disruptors) and lifestyle factors (BMI, alcohol intake, physical activity). Eligible participants had complete genomic, environmental, and behavioral data. Multivariate logistic regression and random-forest algorithms evaluated predictive performance and interaction effects. Prior frameworks (Garcia-Closas et al., JNCI, 2014; Dudbridge et al., Nat Genet, 2018) guided variable selection, data harmonization, and fivefold cross-validation to ensure generalizability across cohorts. Results/Anticipated Results: Models integrating genomic, environmental, and behavioral variables achieved an AUC of 0.87 for breast cancer and 0.83 for ovarian cancer, outperforming PRS-only models (AUC = 0.74, p < 0.001). Fine particulate matter (PM 2.5) exposure and elevated BMI showed significant interactions with BRCA1/2 and CHEK2 variants, amplifying risk by ~1.5–2.3×. Counties with lower environmental quality had 10.8 more breast cancer cases per 100,000 (Gearhart-Serna et al., Sci Rep, 2023). Integrating pollution and lifestyle metrics improved early-risk classification, calibration, and predictive equity across populations, reducing false negatives by 18% and highlighting actionable, low-cost prevention targets for global women’s health. Discussion/Significance of Impact: Integrating genomic, environmental, and behavioral data enhances precision-medicine accuracy for cancer risk assessment. This approach supports equitable prevention by accounting for modifiable exposures and social determinants, reframing cancer prevention through precision public health, and informing population-level screening policy.
Jasmine Alagoz (2026) conducted a cohort in Breast and ovarian cancer (n=5,128). Models integrating genomic, environmental, and behavioral variables vs. PRS-only models was evaluated on Predictive accuracy (AUC) for breast cancer (p=<0.001). Models integrating genomic, environmental, and behavioral variables achieved an AUC of 0.87 for breast cancer and 0.83 for ovarian cancer, outperforming PRS-only models.